As a sweeping economics paper by researchers at the Federal Reserve Bank of Chicago, Forecasting Research Institute (FRI), and numerous top universities found, that attitude may be shifting. As time goes on, top economic experts are increasingly factoring extreme AI disruption into their models. Yet acknowledging a possibility and accepting its inevitable are two very different things — and as the complicated range of sentiments makes clear, an AI jobs apocalypse is still far from certain. The study is a tour-de-force of economic forecasting that surveyed 69 economists, 52 AI specialists, and 38 “superforecasters,” a term for consistently accurate analysts who play the role of “Dune’s” Mentats in the economics world. It found that all three groups expect “significant” progress on AI in the years to come. Forebodingly, the groups all agreed that, as a rule, faster AI progress means lower employment rates overall. On average, economists assigned a 47 percent probability of “moderate“ AI progress by 2030, defined as systems that can operate semi-autonomous research labs, put out high-quality novels, and complete complex projects with oversight.
No comments:
Post a Comment